Reinventing Radio Astronomy – PAF Technology
John O’Sullivan, Centre for Astronomy and Space Science, CSIRO2 April 2013
The origins
• Beginning of time • Optical and later infrared - power detectors/bolometers• Dual/quadruple feed systems for satellite ground station
• 1975? - Ron Ekers and V Radhakrishnan (Groningen) debate whether focal plane has all information
• Interferometer vs lens at focus• 1978 or so - Ron Ekers on sabbatical at CSIRO tries to interest
antenna engineers in fully sampling focal plane
Slide 2AvA2013 Symposium
Still originating
• 1983 – Arnold van Ardenne does heterodyne multibeam mm receiver study
• 1987 – NRAO 7-beam 5.85 GHz receiver• 1987 – Arnold van Ardenne starts work on 350 GHz array for
JCMT• 1988 – NRAO 8-beam Schottky mixer 230 GHz receiver
Slide 3AvA2013 Symposium
And more originating
• 1988 – Cornwell and Napier publish on theory of focal plane coherence to correct aberations, distortions etc.
• 1988 - Ron Ekers joins CSIRO and tries again with MMIC designers for AT Compact Array – runs into shaped dual reflectors!
• 1993 – Trevor Bird and Geoff Poulton multi-feed onboard satellite “illuminator” fo (CSIRO + Hughes)
• 1993 – Parkes 14 beam 21 cm receiver (Bird, Ekers, Stavely-Smith, …)
• 1994 – Arnold van Ardenne et al publish 350 GHz linear array for JCM telescope
• 1995 – Conference on multi-feed systems for radio telesopes at NRAO
Slide 4AvA2013 Symposium
Toward SKA
• 1995 – AvA drives research into aperture array for SKAI in Astron
• 1996 – Rick Fisher paper on fundamentals of phased array feeds on parabolic reflectors
• 1996 – Delft SKAI conference• AvA pushes dense aperture array based on growth projections of
computing/processing. • Ron Ekers raises idea of concentrator plus array in his summary.
• 1997 – 1KT Technical Workshop Sydney• Ron Ekers refers to focal plane arrays in intro talk (plus more
detailed talk later)• Harvey Butcher – arrays and focal plane arrays as Dutch focus• Arnold van Ardenne – comprehensive talk on arrays, station
heirarchy, element types etc.• Rick Fisher – arrays and array beamforming principles.
Slide 5AvA2013 Symposium
First steps
• 2000 – Arnold commissions Vivaldi design with Dan Schaubert (U Mass.)
• 2001 – Arnold and others start Radionet FP5 Faraday program• 2003 – Arnold has sabbatical in Sydney and Marianne Ivashina
(Astron) starts testing at CSIRO of first Astron Vivaldi array tile.• 2001/2 – Peter Hall requests white papers on SKA design
options• Peter Dewdney et al propose large reflector (LAR) with PAF on
aerostat as novel Canadian approach• Ron Ekers pushes concentrator with PAF as Australian approach
– 1D with John Bunton’s cylinder ideas as well as 2D• 2003 – CSIRO embraces PAFs• 2003 – First Astron PAF symposium• 2005 – Peter Hall pushes for distinction between FPA and PAF
nomenclature
Slide 6AvA2013 Symposium
More steps
• 2005 - Stuart Hay proposes connected dipole which soon evolves into a more broadband checkerboard array design (both enthusiastically embraced by myself).
• 2007(?) – Rick Fisher (NRAO), Karl Warnick (BYU) et al propose dipole phased array feed for GBT
Slide 7AvA2013 Symposium
Some working PAFs
• Apertif• 121 element,
• ASKAP• 188 element
• NRAO/BYU• 17 element
• PHAD
Slide 8AvA2013 Symposium
Motivations for PAFs
• Increased FoV• Survey speed
• Simultaneous look directions (transient detection, interference rejection)
• Cost reduction through re-use of dish• Full sampling of image space
• Aberration and reflector distortion, pointing correction• Potential dish cost reduction
• Adaptive interference rejection and spillover reduction• Increased self calibration potential (multiple cal sources in FoV)• Near instantaneous large field imaging
But must hold the line on Aeff/Tsys!
Slide 9
SAeff
Tsys
2
AvA2013 Symposium
Challenges and key technologies
• Competitive Tsys• Challenges in cooling – radiative loss P = σΑΤ4 ~500 W/m2
• Feasible at 5 GHz and above – PHAROS (Glynn et al)
• Also NRAO-BYU dipole to cryogenic receiver• Micro-coolers for receiver? (eg. Schreuder & bij de Vaate)• Uncooled receiver technologies became main game for 21 cm• Understanding and control of all noise contributions key
Slide 10
Fig. 2 Vivaldi Antenna inside Cryostat with Radome (RF
transparentinsulation removed)
AvA2013 Symposium
Array challenges
• Coupling to free space• Low scattering off array• Element beam requirements (pol too)• Influence of inter-element coupling• Scan blindness and excitation of undesired modes• Sampling requirements
• Coupling to receivers• Noise and power matching• Influence of inter-element coupling
• Broad bandwidth designs and matching• Beamforming requirements• Detailed element design and EM+receiver modelling• Measurement and verification• From modelling,measurement to insight and improvement• Cost!!!
• Much work on arrays had apparently been done but it was classified!Slide 11AvA2013 Symposium
The modelling challenge
• Commercial simulators – eg CS Microwave Office and Studio, …
• Very computationally intensive• Accurate (in the right hands)
• Astron and CSIRO increasingly pursued custom simulation• Need to combine EM with electronic to simulate system• Many involved (Ivashina, Maaskant, Woestenburg – Astron)(Hay,
Kot, Christophe, O’Sullivan – CSIRO, Mittra – Penn State U, Craeye - ???)
• CBFM – provides huge numerical gain
Slide 12AvA2013 Symposium
The measurement challenge
• Large arrays with wide angle coverage are challenge to accurately measure
• Astron and CSIRO both have near field antenna ranges• Array as aperture with receivers
• “Popcorn” box/shield with absorbing lid – Astron/NRAO/CSIRO• Moveable absorber – CSIRO• Full spectral cross correlation all ports!
Slide 13AvA2013 Symposium
The measurement challenge(2)
• Array at focal plane• WSRT 25 m – single and interferometer• Greenbank 20 m• Parkes 12 m and 64 m interferometer• Mileura 12 m single dish and 3 element interferometer
Slide 14
Apertif 2008 Sensitivity – compositeBeam Parkes 2012
AvA2013 Symposium
The measurement challenge (3)
• LNA measurement – differential and/or impedance not 50 Ohm• Need all signal and noise parameters to fully characterise array
behaviour• Belototski et al and Astron using tuner-based system to move
source impedance over Smith chart• CSIRO (Shaw) multiple cooled source impedances
Slide 15AvA2013 Symposium
The insight bit
• An example noise matching:• Power transfer and optimum noise match properties first
investigated by brute force modelling• Maaskant and Woestenburg (2007) come up with active
reflection coefficient for beamformed array.
Slide 16AvA2013 Symposium
Further insights
• Hay produces a nice concise formulation which shows fundamental extra array coupling contribution to receiver noise
• Ivashina et al extend active array reflection to equivalent system formulation which allows quantification of various efficiencies
• Infinite array approx allows easy Fourier domain view of noise, beamforming etc.
Slide 17
Tsys Tmin T0
wT Lnarray Lspill N
Gopt
YD YoptI YD YoptI H
w
*
wTw*GD
AvA2013 Symposium
Astronomical
Slide 18
C286 – Apertif 3 interferometersingle PAF (2010)
PKS 1610-771, PKS 1606-772,PKS 1549-790 and PKS 1547-795
ASKAP – 3 element interferometer(2103)AvA2013 Symposium
Further challenges
• PAF• Lowest noise vs cost and power• Vivaldi vs Checkerboard vs ??(bandwidth, losses, cost)• Micro-cooling?
• Signal distribution, beamforming and correlation• Cost, power, size
• Processing and imaging• Calibration, self-cal and automatic calibration of FoV response
Slide 19AvA2013 Symposium
Other Applications
• Satellite TV - Linear Signal 2013• Mobile satellite telecoms• Medical applications – Breast cancer screening
Slide 20AvA2013 Symposium